CRD diagnosis can be confused with other genetic hypocholesterolemias characterized by decreased LDL-cholesterol (LDL-C), such as ABL or homozygous HBL, or with acquired disorders associated with low high-density lipoprotein-cholesterol (HDL)-C. Figure 1 summarizes the differences in familial hypocholesterolemia with low LDL-C. The major clinical and biological manifestations as well as complications are summarized in Table 1 and steps to diagnose CRD are presented in Table 2.

In CRD, consanguinity is frequent but there is usually no noted intrauterine growth retardation. Digestive symptoms are most prominent at the beginning of life. Nonspecific malabsorptive diarrhea is constant and begins in infants shortly after birth. Other digestive symptoms, such as vomiting or abdominal swelling, are often present. They get better within a few days or weeks with a low-fat diet. The long-term evolution of digestive symptoms is variable. The intensity of symptoms decreased over time independent of the level of fat in the diet in some previously reported patients 22. However, our data do not support the hypothesis of intestinal adaptation during CRD. Even though our patients' gastrointestinal symptoms improved on a low-fat diet, diarrhea began again when fat was reintroduced, even in adult patients. Furthermore, there was no improvement in steatorrhea after an average of five years of follow-up 30.

Hepatomegaly is reported to occur in about 20% of CRD patients. Hepatic steatosis is a well known complication of HBL. However, in CRD, at ultrasonography hepatomegaly and steatosis were detected in a few cases. A moderate degree of macrovesicular steatosis was previously reported 2227, but, to our knowledge, no case of cirrhosis has been reported in CRD, in contrast to ABL and HBL 313233. Interestingly, during follow-up, we did not find any correlation between the level of hepatic enzymes and steatosis or hepatomegaly. Serological testing alone is therefore insufficient to detect hepatic morphological problems.

Neurological, muscular or ophthalmic manifestations may raise an alarm in those older patients with a delayed diagnosis as poor compliance to treatment.

The neurological abnormalities described in CRD children are: areflexia at ten and 13 years in a French publication and 13 and 18 years in Strich's publication; areflexia combined with proprioceptive abnormalities at ten and seven years in a Canadian study 212729. Furthermore, more severe neurological degeneration, such as ataxia, myopathy and sensory neuropathy, has been reported in CRD adults 21 and 55 years old 2028. The mean age in the literature of children with clinical neurological manifestations is 12 years, which is significantly older than that for asymptomatic patients (four years). In our cohort, 4/16 patients had electromyographic abnormalities (electromyography with a reduction in sensory nerve conduction velocity and decreased sensory nerve action potential amplitudes). The youngest was 6.5 years old and only one had areflexia as early as age 11. Vitamin E status plays a pivotal role in neurological degenerative complications 3435. In our study, it is relevant to note that the patients with the more pronounced abnormalities had also the lowest vitamin E levels at diagnosis.

Muscular pain and cramps are rare, but classical complaints have been reported by patients. Recently, cardiomyopathy with a decrease in ejection fraction to 40% (normal > 60%) has been described in adults with CRD 15, but its prevalence is unknown. In our cohort, all patients had normal echocardiography, and no signs of cardiomyopathy were found before 23.5 years of age. Muscle histology did not reveal significant lipid accumulation or necrosis, but rather only mild and nonspecific abnormalities: a mild increase in lipid content, focal disorganization of the Z line with the loss of a few Z lines in specific regions of the muscle specimen (personal data), and a variation in muscle fiber diameter with numerous lobulated fibers 15.

With regard to ophthalmic complications, minimal visual abnormalities were previously been reported: micronystagmus, mild deficit in the perception of the blue-yellow axis, and delayed dark adaptation 27. In our populations, ophthalmic complications were mild, evidenced only through functional testing (abnormal visual evoked potentials characterized by waves of higher amplitude and increased latencies, as well as abnormal scotopic electroretinograms). In conclusion, in children, the absence of severe neurologic impairment and retinopathy is compatible with a diagnosis of CRD.

Poor mineralization and delayed bone maturation have previously been observed in CRD 21, probably as a consequence of malabsorption, malnutrition and vitamin D deficiency.

Interestingly, Sar1b GTPase is expressed not only in the intestine but also in the liver, muscle and brain. The same mutation may be involved in different tissue-specific failures, for example, hepatocyte transport of nascent VLDL from the RE to the Golgi thanks to a Sar1b-dependant mechanism 36. We can hypothesize that this dysregulation may be involved in the hepatic steatosis described in CRD. Furthermore, the myocardium has been shown to express the Sar1b gene and secrete apo B lipoproteins 3738. This suggests that the recent description of cardiomyopathy in CRD may be related to the tissue-specific expression of abnormal Sar1b protein. Finally, mild musculoskeletal and neurological abnormalities could also be related to impaired Sar1b in various systems 711.

Hypocholesterolemia associated with chronic diarrhea is common and nonspecific. However, CRD may be suspected with a particular lipid profile. First, the more than 50% decrease in total cholesterol (1.49 ± 0.56 in CRD vs 3.73 ± 0.80 mmol/L in control patients), LDL-C (0.69 ± 0.38 in CRD vs 2.33 ± 0.64 mmol/L in control patients) and HDL-C (0.46 ± 0.08 in CRD vs 1.07 ± 0.22 mmol/L in control patients) in the presence of normal triglycerides (TG) (0.73 mmol/L) is almost pathognomonic 3039. In contrast, both AB and homozygous HB have hardly any measurable TGs much lower total cholesterol levels and no LDL-C. Interestingly, in CRD, the intense hypocholesterolemia is not associated with consequences as severe as in the AB or HB homozygotes. First, the dyslipidemia is much less severe than in AB and HB. Secondly, in CRD, the chylomicrons are not secreted in the classic COP II way, but are built into the RE. Intestinal biopsies show chylomicron-like particles in membrane-bound compartments, reminiscent of the vesiculated channels of the smooth ER and in big membrane-bound compartments 19. We can, therefore, presume that some fat is transported via LDL particles throughout the day and hardly detectable during the 5-hour fat loading test. However, at the puncture, no data support this possibility.

CRD and AB are recessive autosomal diseases, unlike HB which is a dominant form of hypocholesterolemia. Therefore, parental lipid screening may clarify the diagnosis. The absence of hypocholesterolemia in the two parents favors CRD.

Interestingly, the measurement of creatine kinase (CK) may orient towards the diagnosis. Recently and for the first time, muscular abnormalities have been described in CRD patients 15. In our cohort, CK levels were elevated (5 × N) in all patients in whom the measurement was made at diagnosis, except one. The highest level (10 × N) was found in a patient who had the most severe level of neurological impairment (areflexia and abnormal electromyography). However, the CK level does not always correlate well with the severity of the neurological impairment. In conclusion, an elevated level of CK in a patient with hypocholesterolemia can suggest CRD.

Acanthocytes can be seen in advanced liver disease, rare vitamin E genetic deficiency or neurodegenerative syndromes (McLeod), AB, or homozygous HB. In CRD, acanthocytosis is rare, and sometimes transient. It is dependent on the vitamin E level which is a key to red cell membrane integrity. Therefore, the presence of acanthocytosis is hardly compatible with CRD.

As previously mentioned, hepatic abnormalities may be present. Nonspecific hepatic cytolysis is very frequent, but moderate, (1.5-3 × N), and poorly correlates with steatosis and/or hepatomegaly.

Malabsorption of fat leads to deficiency of fat soluble vitamins and essential fatty acid (EFA). Steatorrhea at the time of diagnosis (usually a few months after the onset of the first symptoms) is invariably present and depends on the quantity of fat ingested. With a low-fat regimen, steatorrhea may be falsely negative. The calculation of % fat absorption on a normal fat diet is a key diagnostic test in all patients suspected of CRD.

Fat malabsorption alters fat-soluble vitamins in variable ways. Vitamin E is the most affected among the liposoluble vitamins in CRD, because its transport is highly dependent on apo B-containing lipoproteins 40. Vitamin E deficiency is constant at diagnosis in young children. In CRD, the decrease is drastic and most of the time permanent, even with vitamin E supplementation. Vitamin A is often decreased, but easily corrected with oral supplementation. Finally, vitamin D or K insufficiency detected in half the patients can also be corrected with supplements.

Fat malabsorption also affects the EFA profile. At diagnosis, our CRD patients had abnormal plasma fatty acid profiles, as evidenced by n-6 deficiency (linoleic acid; C18:2n-6) and, surprisingly, normal omega-3 levels. In addition, EFA deficiency manifested by a high 20:3n-9/20:4n-6 ratio, (0.04 ± 0.02 in CRD vs 0.01 ± 0.005 in control patients); normal value <0.02 was observed in the Canadian cohort 30.

The oral fat loading test helps evaluate intestinal fat absorption. Patients fast for 12 h and ingest 50 g of fat per 1.73 m² surface area (flavored commercial cream). Plasma lipids are measured at 2, 3 and 5 h following the fat meal. All patients with CRD do not respond to an oral fat loading test: TGs are normal or slightly decreased at baseline and do not increase at 3 and 5 h while low HDL-C stays low and no chylomicrons are identified. However, this test does not discriminate CRD from the other genetic disorders with hypocholesterolemia.

Upper endoscopy reveals a white duodenal mucosa with normal esophageal and normal gastric mucosa (Figure 2). To increase endoscopic sensitivity, a TG-rich diet is begun three days before the exam to improve the fat loading of enterocytes.

Histology shows multi-vacuolated enterocytes in intestinal villi of normal architecture (Figure 3). However, some patients demonstrate mild atrophic villi 21, which may be confused with celiac disease if lipid vacuoles are not identified. Furthermore, in the same patient, villi are affected to a variable extent and, among different patients, the percent of lipid-laden villi and the region of affected villi can vary substantially 19. In most cases, the vacuolization is seen only in the upper one-third of the villi.

Finally, identification of the SAR1B gene mutation confirms the diagnosis. Since the discovery of SAR1B as the gene that causes CRD 8, 14 different mutations in about 30 patients have been described (Figure 4) 8111641. Only two families have had frame shift mutations that cause CRD. Until now, missense mutations have represented the majority of SAR1B mutations. We previously reported the absence of a correlation in genotype-phenotype in a cohort of 16 patients with CRD 30 and Charcosset et al. analyzed the putative consequences of different mutations including frame shift and missense mutations in CRD. Truncating variants would be expected to induce more severe phenotypes than just missense variants. However, the most deleterious mutation of lipid levels or growth reported in our cohort was a missense mutation (409G > A) 30. Furthermore, the variation in clinical expression with the same mutation in different families 11 and even in the same family with the very surprising asymptomatic homozygote CRD mother, described by Cefalu et al 41, suggests that modifier genes may modulate the transport of the COP II vesicles and that CRD expression is more complex than a simple autosomal recessive disorder.

Follow-up should be directed toward monitoring nutrition and growth, compliance to treatment and the presence of complications involving the liver, the neuromuscular system including the retina as well as bone health. Table 3 proposes a long term management program proposed for patients with CRD.

Growth is a pediatric-specific goal in patients with malabsorption syndromes. With early diagnosis and treatment, catch-up growth can be expected. However, seven of our patients with a delayed diagnosis did not even reach the 20^th percentile of their predicted growth potential of the 40^th percentile. Tracking the growth curve is essential in the follow-up of CRD children. Annual evaluation is appropriate.

The evolution of hepatic cytolysis seems favorable. However, the lack of very long-term data means that patients should be followed carefully. Therefore, measurement of hepatic transaminase levels and ultrasonography should both be carried out. An evaluation every three years after the age of ten seems a reasonable frequency for these noninvasive low-cost exams. Liver stiffness measured by elastometry (Fibroscan) could be a useful tool to detect early fibrosis, but this proposal needs further evaluation.

Neurological, muscular and ophthalmologic complications in CRD are less severe than in ABL or homozygous HBL, but may occur during infancy and require rigorous follow-up and compliance with vitamin treatments. The lack of data on long-term outcome in treated patients with CRD means that rigorous follow-up is required. Although high CK levels may indicate severe muscular impairment, they are inadequate to evaluate neuromuscular progression. Therefore, beginning at pre-puberty (ten years old), neurological and ophthalmologic examinations may be necessary every three years to detect complications such as developmental retardation, areflexia, ataxia, dysarthria, loss of deep proprioception, muscular weakness and decreased night vision. Furthermore, it seems reasonable to schedule a specialized neurology and ophthalmology consultation every three years to conduct electromyography, measure visual evoked potentials, and perform electroretinography, because electrical abnormalities occur before clinical symptoms. Finally, we suggest echocardiography in adulthood (> 18 years), since early detection of cardiac impairment may motivate patients who have discontinued their vitamin therapy to restart it. It may also help to schedule treatment before clinical manifestations of cardiac insufficiency develop. A frequency of every three years seems reasonable for this noninvasive exam.

Dual-energy X-ray absorptiometry (DEXA) may be performed to measure the bone mineral content of the total skeleton, the best index of bone mineralization in children 4243.

Table 4 summarizes our proposed treatment protocol. A low-fat diet containing an appropriate amount and ratio of n-6/n-3 is necessary to improve digestive symptoms. In very young children, milk preparations with medium-chain TG may improve diarrhea and correct malnutrition within a few days but tolerance can be a problem. In older children, a regimen low in long-chain fatty acids is usually sufficient to decrease symptoms.

Vitamins E and A are implicated in neurological, muscular and ophthalmic complications. High dosages of vitamin E (100 IU/kg/d) have been reported to prevent, slow or improve neurological complications in different hypocholesterolemia disorders such as ABL 354445464748. Furthermore, these high intakes are safe. Relatively few side effects in the short term have been reported in double-blind studies of vitamin E at doses as high as 2000 or 3200 IU/d 49. Alpha-tocopherol either in aqueous or lipid form is the most effective form of vitamin E to prevent neurological complications. Indeed, nerve tissue is characterized by slow exchange and high selectivity for alpha-tocopherol 50. Lipid malabsorption is a common feature of several diseases, such as celiac disease, cystic fibrosis and cholestasis. In the attempt to promote liposoluble vitamin absorption, new pharmacologic formulations of vitamin E have been developed. Tocopheryl polyethylene glycol succinate is a hydrosoluble form of vitamin E. This water-soluble formulation showed a marked and statistically significant increase in the absorption of gamma-tocopherol in malabsorbing patients with cystic fibrosis (CF) compared with a classic oil-based formulation 51. Recently, intestinal absorption of water-soluble vitamin E (tocofersolan) was compared with a water-miscible form in 12 CF or cholestatic children 52. Oral tocofersolan was more bio available than the water-soluble formulation in children with chronic cholestasis and similarly bio available in CF. This suggests that water-soluble vitamin E may represent an alternative to painful intramuscular vitamin E injections in chronic cholestasis, or other oral formulations in CF. However, the mechanism responsible for fat malabsorption in CRD concerns the absorptive phase as opposed to the digestive phase in CF and cholestatic syndromes. To our knowledge, no specific studies with these new vitamin E preparations have been conducted in CRD. For CRD, further investigation on the efficacy of water-soluble vitamin E is needed. Our 16 CRD patients received an average of 51 ± 32 IU/kg/d of vitamin E with some receiving a dosage as high as 100 IU/kg/d. Despite higher dosages of oral vitamins E and A and, in some cases, compliance ensured by intravenous administration, four of the seven French patients developed mild clinical signs of neurological impairment. This may be due to the delayed diagnosis or influence of the mutation. However, the majority of patients (12/16 patients) with a medium dosage of 50 IU/kg/d, sometimes as low as 20 IU/kg/d, did not develop clinical or electrical neuro-ophthalmologic complications when the treatment was given during the first years of life. Therefore, it seems reasonable to prescribe only 50 IU/kg/d of vitamin E to patients with CRD if the disease is diagnosed during the first year of life.

To prevent complications, our data do not suggest that the intravenous route is warranted. The oral route is sufficient, as illustrated by the nine patients in the Canadian cohort who were treated orally early on and who remained free of neurological complications (including two patients aged 22).

Determination of the appropriate dosage of vitamin E can be problematic. In our cohort, plasma concentration measurements were too imprecise to serve as a guide for clinicians. Plasma levels after supplementation reached only one half to two thirds of normal levels, independent of the vitamin E dose. Among the 16 patients with 20, 40 or 100 IU/kg/d, one may find the same plasma vitamin E levels of around 60% of the lower normal range. This level is comparable with that obtained in ABL after oral supplementation with high doses (45% of the normal range for 100 IU/kg/d of vitamin E) 53. Furthermore, plasma levels of vitamin E did not necessarily correlate with vitamin E dosages. The severity of steatorrhea was not correlated with vitamin E and A deficiency in our patients. Therefore, vitamin treatment cannot be adjusted to the severity of fat malabsorption. An interesting alternative could be to measure vitamin E levels in subcutaneous adipose tissue. It was previously demonstrated that after supplementation, adipose tissue concentrations reach normal levels, even if plasma levels remain low 54.

Experience with our 16 patients suggests that vitamin A at a dosage of 15000 IU/d is effective in combination with vitamin E to prevent ophthalmologic complications in CRD.

When vitamin D treatment (800 to 1200 IU/d) is started early, it prevents osteopenia, as illustrated by the nine Canadian patients.

Vitamin K was administered at the same dosage of 15 mg/week in the two groups. The plasma values were not measured, but coagulation screening was normal and no hemorrhages were detected in any patient during the entire follow-up.

Patients with lipid malabsorption are at an increased risk for EFA deficiency. Such as mentioned previously, a low-fat diet is necessary to improve digestive symptoms, but it may increase the risk of EFA deficiency. Therefore, dietary counseling is needed not only to decrease fat intake but also to maintain sufficient caloric and EFA intake. It was recommended that the nine Canadian patients add one to three teaspoons per day of soybean oil to meals and increase fish consumption to two to three times per week. The objective was to reach 3-5% of total calories with omega 6, and 0.5-1% with omega 3. The French cohort (seven patients) had the same dietary recommendations, but each month four of them also received an infusion of 80 g of lipid emulsion containing 8 g of EFA (linoleic, 80% olive oil and 20% purified soy oil). The other three patients drank 25 g per day of oral supplements containing medium-chain TG (Liprocil, 80% MCT, Nestlé, France). Intravenous and oral supplementations were unable to normalize the EFA plasma levels. Furthermore, the clinical impact of this supplementation is not obvious, since the Canadian cohort had less supplementation and worse plasma levels, but fewer complications. This discrepancy may be explained by a difference between plasma fatty acid levels and the fatty acid concentration in tissues, since dietary fatty acids are mainly carried by chylomicron vesicles, which are dramatically decreased in CRD. Adipose tissue biopsies could provide some answers to this question. Simple counseling to increase oral EFAs with vegetable oils and fish seems appropriate in CRD, as illustrated by the excellent clinical progression of the nine Canadian patients despite the low plasma omega-6 levels. However, prospective clinical studies with fatty acid supplementation are needed to evaluate the optimal dose, route of administration and long-term impact of EFA treatment. Similarly, the role of medium-chain TG supplementation in CRD patients remains to be defined.